The machines bought Super Bowl airtime and we rank them
Super Bowl LX proved that AI companies have money, ambition, and wildly different ideas about what they're selling us.
Twenty-three percent of Super Bowl LX commercials featured artificial intelligence. Fifteen spots out of sixty-six. By the end of the first quarter, fans on X were already exhausted. The crypto-bro era of 2022 has found its successor. This one has better PR.
But unlike the parade of indistinguishable blockchain pitches from years past, the AI ads told us something. They revealed, in thirty-second bursts, which companies understand what they’re building and which are still figuring out how to explain it to 120 million people eating guacamole.
The results split cleanly. One company made art. One made a promise it probably can’t keep. One made an argument. And one made whatever Chris Hemsworth was doing with that snake.
The Winner: Anthropic’s “A Time and a Place”
Anthropic spent $8 million on four spots. Every dollar shows.
The campaign, directed by Jeff Low and set to Dr. Dre’s “What’s the Difference,” depicts AI conversations gone wrong. In “Betrayal,” a man asks a chatbot how to talk to his mother. The AI (played by a blonde woman with the vacant warmth of a wellness influencer) offers helpful advice, then pivots mid-sentence to hawk a cougar-dating site called Golden Encounters. The man’s face, baffled and betrayed, is the whole joke. The tagline lands: “Ads are coming to AI. But not to Claude.”
The other spots follow the template. A nervous entrepreneur gets mentor-y guidance that swerves into a payday-loan pitch. A scrawny guy asks about six-pack abs and gets served an ad for height-boosting insoles. Each one earns its title: “Treachery,” “Deception,” “Violation.”
Tom Warren of The Verge called them “hilarious.” One X user declared Anthropic had “buried OpenAI” with no coming back. Even Sam Altman admitted he laughed, then called them “clearly dishonest.”
What makes these ads work isn’t the humor. It’s the clarity. Anthropic knows exactly what it’s selling: trust. In a world where AI is about to get very intimate with your search history, your health questions, your midnight anxieties, the absence of advertising becomes a feature worth paying for.
Felix Richter, Mother’s global chief creative officer, put it plainly: “We’re using advertising’s biggest stage to ask a simple question: Does it belong everywhere?”
The answer, at least in Anthropic’s telling, is no.
The Runner-Up: Google’s “New Home”
If Anthropic made an argument, Google made a feeling.
“New Home” opens on a mother and son facing the particular dread of moving day. Boxes everywhere. Empty rooms. The kid is anxious. So the mother pulls out her phone and asks Gemini to show her what his room could look like: his toys scattered on the floor, the walls painted his favorite shade of blue. The transformation is gentle, almost magic-realist. Randy Newman’s “Feels Like Home” does the emotional work.
The spot, directed by Daniel Mercadante and created by Google Creative Lab, never feels like a tech demo. Gemini operates in the background, enabling rather than performing. It’s the first AI ad I’ve seen that understands a basic truth about technology: the best tools disappear.
AdWeek called it “sweet but predictable.” Fair. The ad isn’t surprising. But it’s visually the most accomplished of the bunch: warm color grading, confident pacing, no desperation. Google knows how to make a commercial. The question is whether anyone remembers it’s about Gemini and not just about a nice mom.
The Dodge: OpenAI’s “Builders”
OpenAI returned to the Super Bowl with a 60-second spot about, according to Sam Altman, “builders, and how anyone can now build anything.”
The ad follows a thread of human curiosity: classroom notebooks, early code, a Linux DVD, people working alongside ChatGPT to create apps and websites. It’s aspirational. All OpenAI marketing is aspirational. The tool appears at its most flattering, surrounded by earnest strivers achieving their dreams.
Production values are solid. The company leaned on Sora and ChatGPT for ideation, though the final cut features real people. Doomsday Entertainment produced; MakeMake edited. It looks expensive because it was. OpenAI spent roughly $14 million on its Super Bowl buy.
Here’s the problem: the ad is about Codex, OpenAI’s coding assistant, while the company is simultaneously testing ads in ChatGPT. The inspirational story of builders building feels dishonest when the company is also building an ad-delivery system for its free tier. Altman insists the ads won’t work the way Anthropic depicts, but the company’s official positionacknowledges ads will be “conversation-specific.” That’s precisely what makes Anthropic’s parody land.
FleetingBits on X offered the sharpest critique: the Anthropic ads were “designed for the SF Twitter bubble,” assuming America already views Claude as the trustworthy alternative. Maybe. But OpenAI’s spot assumes something equally shaky: that we’ll keep believing the company is about empowerment while it figures out how to monetize attention.
One X user captured the tension: “Ads give OpenAI a Meta-like multiple. Anthropic’s focus on monetizing API tokens and enterprise seats will give them airline-to-grocery-store multiples over time.”
Translation: both companies are making bets. Anthropic is betting on premium. OpenAI is betting on scale. The Super Bowl just made those bets public.
The Disaster: Amazon’s “Alexaaaa+”
I watched this one twice. I’m still not sure what happened.
Chris Hemsworth plays himself, convinced that Alexa+ is trying to kill him. He rattles off scenarios: the AI closes the garage door on his head, traps him underwater in the pool, summons a bear, explodes the fireplace. His wife, Elsa Pataky, watches with bemused patience. In the end, Alexa offers to book him a massage. The tagline is “Scary Good.”
The problem isn’t Hemsworth. He commits to the bit. The problem is the bit. An ad whose central joke is “our AI might murder you” is a strange choice for a product Amazon wants people to invite into their bedrooms. Mike Stanton on Xnailed it: “This is funny but also an effective ad AGAINST Alexa.”
This is funny but also an effective ad AGAINST Alexa. https://t.co/QXM99nPRpN
— Mike Stanton (@projomike) February 6, 2026
Entertainment Weekly described Hemsworth as “gaslighting us about AI.” TVLine ranked the spot among the worst of the night, noting that “the payoff isn’t big enough, and so you’re just left wondering if Alexa will actually start sabotaging things around your home.”
The 90-second extended cut is worse. Hemsworth gets hurled against a tree, attacked by a snake, yanked from his workout. It’s dark comedy, but comedy requires landing somewhere. This one just keeps going, hoping volume substitutes for structure.
Amazon wanted to acknowledge AI anxiety and defuse it with humor. Instead, they made a commercial that confirms the anxiety while demonstrating nothing about why Alexa+ is worth $19.99 a month.
The One We Missed: Microsoft’s Copilot
Microsoft ran a spot too. It was about NFL statistics.
Two recruiters on a sideline ask Copilot to chart linebacker prospects with 40-yard times under 4.6 seconds. The AI complies. They drill deeper: which ones have leadership skills? Which might get 100 tackles? Bar charts appear.
It’s fine. It’s forgettable. AdWeek noted that Microsoft “showed Copilot’s power but not its personality.” The spot exists to remind you that Microsoft is still an NFL sponsor.
In a Super Bowl dominated by AI companies fighting over the future of how we interact with machines, Microsoft made a commercial about spreadsheets. It’s honest. Most enterprise AI actually does this. It’s also the ad equivalent of bringing a calculator to a knife fight.
What the Ads Tell Us
The AI industry spent Sunday introducing itself to America. The introductions varied.
Anthropic said: We’re the ones who won’t sell you out. Google said: We’ll make moving day slightly less sad. OpenAIsaid: Anyone can build anything (and soon, you’ll see ads while you do it). Amazon said: Our AI might kill you but probably won’t. Microsoft said: Charts.
Only one of these is a coherent brand position. Anthropic’s gamble (that trust is worth more than scale) may prove naive. Dario Amodei is projecting $70 billion ARR by 2027, profitability by 2028. Those numbers require a lot of people to choose the premium option.
But the Super Bowl isn’t about business models. It’s about storytelling. Anthropic told the best story: AI is about to get very personal, and you should think carefully about who you let into the conversation.
The others told us they exist. That’s not nothing. It’s just not enough.
Watch the Ads
| Company | Spot | Link |
|---|---|---|
| Anthropic | “Betrayal” | iSpot |
| Anthropic | “Violation” (Six-Pack) | iSpot |
| “New Home” | iSpot | |
| OpenAI | “Builders” | iSpot |
| Amazon | “Alexaaaa+” | iSpot |
| Microsoft | “NFL: Turn Raw Data Into Insights” | iSpot |
CO/AI covers the intersection of AI technology, business, and whatever the machines are up to this week.
Recent Blog Posts
Six ideas from the Musk-Dwarkesh podcast I can’t stop thinking about
I spent three days with this podcast. Listened on a walk, in the car, at my desk with a notepad. Three hours is a lot to ask of anyone, especially when half of it is Musk riffing on turbine blade casting and lunar mass drivers. But there are five or six ideas buried in here that I keep turning over. The conversation features Dwarkesh Patel and Stripe co-founder John Collison pressing Musk on orbital data centers, humanoid robots, China, AI alignment, and DOGE. It came days after SpaceX and xAI officially merged, a $1.25 trillion combination that sounds insane until you hear...
Feb 3, 2026The Developer Productivity Paradox
Here's what nobody's telling you about AI coding assistants: they work. And that's exactly what should worry you. Two studies published this month punch a hole in the "AI makes developers 10x faster" story. The data pointssomewhere darker: AI coding tools deliver speed while eroding the skills developers need to use that speed well. The Numbers Don't Lie (But They Do Surprise) Anthropic ran a randomized controlled trial, published January 29, 2026. They put 52 professional developers througha new programming library. Half used AI assistants. Half coded by hand. The results weren't close. Developers using AI scored 17% lower on...
Feb 3, 2026The Lobsters Are Talking
January 2026 will be remembered as the week agentic AI stopped being theoretical. For three years, we've debated what autonomous agents might do. We wrote papers. We held conferences. We speculated about alignment and control and the risks of systems that could act independently in the world. It was all very intellectual, very abstract, very safe. Then someone open-sourced a working agent framework. And within days, thousands of these agents were talking to each other on a social network built specifically for them while we could only watch. I've been building things on the internet for over two decades. I...